9 April, 2002 PROC. ENTOMOL. SOC. WASH. 104(2), 2002, pp. 437-446 CONSTRUCTION OF A DEFENSIVE TRASH PACKET FROM SYCAMORE LEAF TRICHOMES BY A CHRYSOPID LARVA (NEUROPTERA: CHRYSOPIDAE) Thomas Eisner, James E. Carrel, Eileen Van Tassel, E. Richard Hoebeke, and Maria Eisner (TE, ME) Department of Neurobiology and Behavior, Cornell University, Ithaca, NY 14853, U.S.A. (e-mail: [email protected]; mle3@ cornell.edu); (JEC) Division of Biological Sciences, University of Missouri, Columbia, MO 65211, U.S.A. (e-mail: [email protected]); (EVT) Department of Entomology, Michigan State University, East Lansing, MI 48824, U.S.A. (e-mail: [email protected]); (ERH) Department of Entomology, Cornell University, Ithaca, NY 14853, U.S.A. (e-mail: [email protected]) Abstract. The behavior of a chrysopid larva from Arizona is described, identified as Ceraeochrysa lineaticornis ( Fitch), which constructs its trash packet from the trichomes of sycamore leaves ( Platanus wrightii S. Watson). The trichomes are particularly dense on the underside of the leaves. To fashion the packet, the larva uses trichomes stuck loosely to the leaf, and others that it plucks from the leaf surface. The packet on the mature larva weighs on average 2.2 mg, an equivalent of approximately two leaf undersides worth of trichomes. Evidence is presented indicating that the packet provides the larva with protection against predation. It is argued that the trichomes are defensive in the sycamore tree itself, and that the chrysopid provides yet another example of an insect that benefits from utilization of a plant defense. Interestingly, the chrysopid benefits the sycamore tree itself. Although it usurps the tree's defense, it aids the tree by preying on a specialist herbivore, the tingid, Corythucha confraterna Gibson, which feeds on the sycamore tree unbothered by the trichomes. Key Words: animal defense, Neuroptera, Chrysopidae, Ceraeochrysa lineaticornis, plant defense, trichome, insect-plant interaction. Many chrysopid larvae have the habit of against insectan predators (New 1969, Principi collecting exogenous materials and placing 1946, Eisner et al. 1978). Not all chrysopid them on their backs, forming so called trash larvae are trash carriers, but those that are, are packets that they retain throughout larval life obligatory so, and the priority that larvae give (Smith 1922, Canard et al. 1984). The materials to forming trash packets has a significant they use to fashion the packets are variable and genetic component (Milbrath et al. 1993, may consist of vegetable matter, arthropod Tauber et al. 1995). remains, insect waxes, or general debris (Smith Here we describe the behavior of a 1922, Slocum and Lawrey 1976, New 1969, chrysopid larva that constructs its packet from Eisner et al. 1978, Canard et al. 1984). Existing trichomes that it takes from the leaves of evidence indicates the packets act as physical sycamore trees (Platanus wrighii S. Watson) in shields that provide the larvae with protection Arizona. We present data on the feeding habits 438 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON VOLUME 104, NUMBER 2 439 of the larva, as well as on how the packet is that were tingid-infested. They took readily to constructed and used in defense. Our the tingids and appeared to feed on the observations were made in the hamlet of nymphs only. Maintained on this diet many of Portal, Cochise County, Arizona, in 1967 the larvae went on to pupate and develop into (September) and 1982 (August). We refrained adults. As is typical for trash-carrying from publishing earlier because we had failed chrysopids, larvae retained the trash packet as to obtain a definitive identification of the an outer covering of their cocoon when they chrysopid. This problem has now been pupated. The cocoons therefore have the same remedied, thanks to the courtesy of Catherine appearance as the larvae, except that they are and Maurice Tauber, who have informed us non-ambulatory. In the field we found several that the chrysopid is Ceraeochrysa lineaticornis cocoons on the trunks of sycamore trees. Most (Fitch) (voucher specimens have been larvae that we used in our experiments were deposited in the Cornell University Insect probably in their last instar. Collection). Electronmicroscopy.For examination with the scanning electronmicroscope, specimens MATERIALS AND METHODS (chrysopid larvae, pieces of sycamore leaf) were preserved in the field in 70% ethanol, then critical point dried and gold coated in the Field site. We first noted the larvae on a group of sycamore trees growing on the banks laboratory. Predation tests. Of the predators used, the of Cave Creek, in Portal itself, on the grounds of the Cave Creek Ranch, where we were in reduviid, identified as Pselliopus latispina residence (Fig. 1A). We subsequently found Hussey, is doubtless a natural enemy of the them also on sycamore trees near Cave Creek, chrysopid. It occurred commonly on the leaves up to several miles upstream from the original and branches of the sycamore trees themselves, location. The larvae were readily spotted by and was once actually observed feeding on a C. their white trash packet, which rendered them lineaticornis in the field. extremely conspicuous against the green color The ant used in predation tests, of the sycamore leaves on which they were Pogonomyrmex barbatus (Smith), may not itself found (Fig. 2A). As they scurried about, they be a primary enemy of the chrysopid, but there resembled tiny ambulatory cotton wads. can be little question that ants as such figure Careful scrutiny of the visually accessible among the larva's natural predators. We lower branches of the trees usually revealed routinely observed ants foraging singly on the presence of many larvae per tree. leaves, branches and trunks of the sycamore trees. Maintenance of larvae.Larvae were maintained on freshly clipped sycamore leaves Statistics.Numerical averages are given as in plastic containers of various sizes, including mean ± S.E. Petri dishes. In the field, larvae had been found on repeated occasions feeding on a RESULTS tingid. Corythucha confraterna Gibson, whose colonies were of common occurrence on the The trichomes.Both surfaces of the sycamore sycamore leaves. Captive chrysopids were leaves bear trichomes, but the structures are therefore always provided with some leaves much more densely distributed on the ← Fig. 1. A, Stand of sycamore trees on which the chrysopid larva was discovered (Portal, AZ). B, Close-up of underside of sycamore leaf showing trichomes. C, Underside of sycamore leaf from which the trichomes have been removed; the detached trichomes from the wad seen in the ampule. D, Enlarged view of leaf underside (scanning electronmicrograph). E, Base of a trichome showing the hinge (arrow). F, Base of trichome that has been torn off at the level of the hinge. Reference bars: D = 0.5 mm; E = 20 µm. 440 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON underside (Figs. 1B, D). Individually the arched that region upward and forward every trichomes are usually multiply branched, and time that the head flexed backward to deliver a they are flexibly hinged near their base, where load. Later pluckings, delivered to the more there is a break in the core of the shaft and the accessible anterior regions of the back, were wall is reduced to a thin-walled tube (Fig. 1E, applied without simultaneous postural F). If one pulls on individual trichomes with adjustments of the body. forceps, they tend to detach at the level of these A visual demonstration of the sequence of hinges. They evidently detach spontaneously trichome delivery during packet construction in large numbers from the leaf surface, as was obtained by giving denuded larvae access evidenced by the fact that the underside of to trichomes of different colors. Packets were sycamore leaves usually has a loose coating of removed from a series of larvae and were then detached trichomes. Microscopic examination stained either in red (with acid fuchsin), in verified clearly that these lose trichomes had black (with chlorazol black), or kept unstained. broken off at the level of the hinges. Denuded larvae that were then given access to Packet construction.Removal of the trash teased apart samples of these packets, in the packets from the larvae could be effected easily sequence of black (for 30 min), to unstained by teasing away the trichomes with forceps. (for 20 min), and to red (for 60 min), Eight larvae that were thus denuded and re- constructed packets in which the trichomes released into their leaf-containing enclosures, were laid out in three colored bands, in the commenced reloading almost at once. In sequence red, white, and black from fore to aft typical chrysopid fashion (New 1969, Principi (Fig. 2B). 1940, Smith 1922), using their jaws as a two- Package repair.Use of stained trichomes pronged fork, they scooped up load upon load also yielded visual evidence of the precision of trichomes and placed these upon their backs with which the larvae are able to repair (Figs. 2C-F). Most often they procured damage to their package. Four larvae, bearing trichomes from the clusters of available loose natural unstained packets, were treated as ones, but they also pried many off with their follows: jaws. It seemed clear, moreover, that they were (1) Trichomes removed from center of programmed to utilize sycamore trichomes packet, leaving the larvae with a ring-shaped only. If released in a Petri dish with general shield. particulate debris they tended to ignore such (2) Posterior half of packet removed. matter, but if then offered a sycamore leaf, they (3) Left half of packet removed. usually returned promptly to the task of (4) No trichomes removed (control). reloading. Like trash-carrying chrysopids The larvae were then confined with a supply generally (Canard et al. 1984), the larvae are of black trichomes, and checked for condition equipped with a set of special hooked bristles of their packets after 24 hours.